A Stages of Change (SOC) treatment approach, supplemented by Motivational Interviewing techniques (Miller & Rollnick, 2002), was compared with a standard cognitive-behavioral therapy gender-reeducation (CBTGR) approach for its effectiveness in reducing men's violence against their partners. A total of 528 male batterers from the Montgomery County, Maryland Abused Persons Program were randomly assigned to 49 26-week groups in either the SOC or CBTGR condition. English-speaking and Spanish-speaking groups were conducted in both treatment conditions. While certain differences emerged between English-and Spanish-speakers and their partners, men in the two treatment conditions (and their partners) did not differ with respect to demographics, lifetime violence, psychopathology, multiple admissions, histories of trauma, or source of referral as a function of their assignment to different treatment conditions. Differences between women who were successfully contacted for follow-up and those who were not were minimal, indicating the generalizability of victim follow-up reports.Self-reported aggression at post-treatment was unrelated to treatment condition. However, based on victim follow-up reports, the SOC curriculum was more effective than standard treatment in reducing women's risk for physical aggression, especially among English-speakers. The SOC condition was particularly effective for first-time offenders, for men who were court-ordered to treatment and for men in an earlier underlying stage of change. On the other hand, the SOC condition was no more effective than the CBTGR condition in reducing the violence of men with multiple admissions, suggesting the need for warning their victims and for using a coordinated community response. The SOC condition was also not effective for men who were self-referred to treatment, presumably because it did not quickly provide the anger-management and conflict resolution skills these men were seeking. The two treatments did not differ in rates of attendance (either number of sessions attended or treatment completion), working alliance or group cohesion. Different predictors of attendance did emerge between the two conditions. Moreover, blind ratings of audiotapes of sessions significantly differentiated the two conditions. Specific treatment considerations for Spanish-speakers were explored as was the necessity of attending to the trauma histories of both men and their partners. Finally, the reactions of therapists working with the SOC curriculum were described as were the limitations of this study and suggestions for future research. It was concluded that within the context of these limitations, this project allowed a fairly rigorous but also realistic test of the usefulness of SOC treatment in reducing men's risk for violence.